informatik-vortraege May 2021

informatik-vortraege@lists.rwth-aachen.de

5 participants
7 discussions

Einladung Informatik-Oberseminar Malte Nuhn

by Sekretariat I6

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 12. Juli 2019, 10.00 Uhr Ort: Informatikzentrum, E3, Raum 9222 Referent: Dipl.-Inform. Malte Nuhn Thema: Unsupervised Training with Applications in Natural Language Processing// Abstract: The state-of-the-art algorithms for various natural language processing tasks require large amounts of labeled training data. At the same time, obtaining labeled data of high quality is often the most costly step in setting up natural language processing systems.Opposed to this, unlabeled data is much cheaper to obtain and available in larger amounts.Currently, only few training algorithms make use of unlabeled data. In practice, training with only unlabeled data is not performed at all. In this thesis, we study how unlabeled data can be used to train a variety of models used in natural language processing. In particular, we study models applicable to solving substitution ciphers, spelling correction, and machine translation. This thesis lays the groundwork for unsupervised training by presenting and analyzing the corresponding models and unsupervised training problems in a consistent manner.We show that the unsupervised training problem that occurs when breaking one-to-one substitution ciphers is equivalent to the quadratic assignment problem (QAP) if a bigram language model is incorporated and therefore NP-hard. Based on this analysis, we present an effective algorithm for unsupervised training for deterministic substitutions. In the case of English one-to-one substitution ciphers, we show that our novel algorithm achieves results close to human performance, as presented in [Shannon 49]. Also, with this algorithm, we present, to the best of our knowledge, the first automatic decipherment of the second part of the Beale ciphers.Further, for the task of spelling correction, we work out the details of the EM algorithm [Dempster & Laird + 77] and experimentally show that the error rates achieved using purely unsupervised training reach those of supervised training.For handling large vocabularies, we introduce a novel model initialization as well as multiple training procedures that significantly speed up training without hurting the performance of the resulting models significantly.By incorporating an alignment model, we further extend this model such that it can be applied to the task of machine translation. We show that the true lexical and alignment model parameters can be learned without any labeled data: We experimentally show that the corresponding likelihood function attains its maximum for the true model parameters if a sufficient amount of unlabeled data is available. Further, for the problem of spelling correction with symbol substitutions and local swaps, we also show experimentally that the performance achieved with purely unsupervised EM training reaches that of supervised training. Finally, using the methods developed in this thesis, we present results on an unsupervised training task for machine translation with a ten times larger vocabulary than that of tasks investigated in previous work. Es laden ein: die Dozentinnen und Dozenten der Informatik _______________________________________________ -- -- Stephanie Jansen Faculty of Mathematics, Computer Science and Natural Sciences HLTPR - Human Language Technology and Pattern Recognition RWTH Aachen University Ahornstraße 55 D-52074 Aachen Tel. Frau Jansen: +49 241 80-216 06 Tel. Frau Andersen: +49 241 80-216 01 Fax: +49 241 80-22219 sek(a)i6.informatik.rwth-aachen.de www.hltpr.rwth-aachen.de Tel: +49 241 80-216 01/06 Fax: +49 241 80-22219 sek(a)i6.informatik.rwth-aachen.de www.hltpr.rwth-aachen.de

2 years, 3 months

UnRAVeL "Behind the Scenes" Survey Lecture

by Tim Seppelt

Dear all, part of the programme of the research training group UnRAVeL is a series of introductory lectures on the topics of „randomness“ and „uncertainty“ in UnRAVeL’s research thrusts algorithms and complexity, verification, logic and languages, and their application scenarios. Each lecture is delivered by one of the researchers involved in UnRAVeL. The main aim is to provide doctoral researchers as well as master students a broad overview of the subjects of UnRAVeL. This year, 12 UnRAVeL professors will answer the following questions, based on one of their recent scientific results: * How did you get to this result? * How did you come up with certain key ideas? * How did you cope with obstacles on the way? Which ideas you had did not work out? Following these talks, PhD students will give an informal summary of their doctoral studies within UnRAVeL. All interested doctoral researchers and master students are invited to attend the UnRAVeL lecture series 2021 and engage in discussions with researchers and doctoral students. Details information can be found on https://www.unravel.rwth-aachen.de/cms/UnRAVeL/Studium/~pzix/Ringvorlesung-… All events take place on *Thursdays from 16:30 to 18:00 on Zoom* https://rwth.zoom.us/j/96043715437?pwd=U0dRczkyQjRCY21abW13TDNmUHlhUT09 * 15/04/2021 Survey Lecture: Erika Ábrahám: Probabilistic Hyperproperties * 22/04/2021 Jürgen Giesl: Inferring Expected Runtimes of Probabilistic Programs * 29/04/2021 Erich Grädel: Hidden Variables in Quantum Mechanics and Logics of Dependence and Independence * 06/05/2021 Christof Löding: Learning Automata for Infinite Words * 20/05/2021 Martin Grohe: The Logic of Graph Neural Networks * 10/06/2021 Britta Peis: Sensitivity Analysis for Submodular Function Optimization with Applications in Algorithmic Game Theory * 17/06/2021 Nils Nießen: Optimised Maintenance of Railway Infrastructure * 24/06/2021 Gerhard Lakemeyer: Uncertainty in Robotics * 01/07/2021 Joost-Pieter Katoen: The Surprises of Probabilistic Termination * 08/07/2021 Christina Büsing: Robust Minimum Cost Flow Problem Under Consistent Flow Constraints * 15/07/2021 Ringvorlesung: Gerhard Woeginger: Bilevel optimization * 22/07/2021 Ulrike Meyer: Malware Detection We are looking forward to seeing you at the lectures. Best regards, Tim Seppelt for the organisation committee https://www.unravel.rwth-aachen.de/global/show_picture.asp?id=aaaaaaaaaydoc…

3 years, 1 month

Einladung: Informatik-Oberseminar Andrea Schnorr

by Andrea Schnorr

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 19. Februar 2021, 11.00 Uhr Zoom: https://rwth.zoom.us/j/2452218628 Referent: Andrea Schnorr, M.Sc. LuFG i12 Thema: Feature Tracking for Space-Filling Structures Abstract: Feature-based visualization is a proven strategy to deal with the massive amounts of data emerging from time-dependent simulations: the analysis focuses on meaningful structures, i.e., said features. Feature tracking algorithms aim at automatically finding corresponding objects in successive time steps of these time-dependent data sets in order to assemble the individual objects into spatio-temporal features. Classically, feature-based visualization has focused on sparse structures, i.e. structures which cover only a small portion of the data domain. Given a sufficiently high temporal resolution, existing tracking approaches are able to reliably resolve the correspondence between feature objects of successive time steps. Our research is motivated by our collaborators' work on the statistical analysis of structures that are space-filling by definition: dissipation elements. Space-filling structures partition the entire domain. Our collaborators aim at extending their statistical analysis to a time-dependent setting. Hence, we introduce an efficient approach for general feature tracking which handles both sparse and space-filling data. To this end, we develop a framework for automatic evaluation of tracking approaches, an algorithmic framework for feature tracking, and an efficient implementation of this framework. First, we propose a novel evaluation framework based on algorithmic data generators, which provide synthetic data sets and the corresponding ground truth data. This framework facilitates the structured quantitative analysis of an approach's feature tracking performance and the comparison of different approaches based on the resulting measurements. Second, we introduce a novel approach for tracking both sparse and space-filling features. The correspondence between neighboring time-steps is determined by successively solving two graph optimization problems. In the first phase, one-to-one assignments are resolved by computing a maximum-weight, maximum-cardinality matching on a bi-partite graph. In its second phase, the algorithm detects events by finding a maximum weight independent set in a graph of all possible, potentially conflicting event explanations. Third, we show an optimized version of the second stage of the tracking framework which exploits the model-specific graph structure arising for the tracking problem. The method's effectiveness is demonstrated by a set of case studies including the use of the evaluation framework as well as the analysis of miscellaneous real-world simulation data sets. Es laden ein: die Dozentinnen und Dozenten der Informatik

3 years, 2 months

Einladung: Informatik-Kolloquium Holger Hoos

by Martin Grohe

********************************************************************** * * * Einladung * * * * Informatik-Kolloquium * * * ********************************************************************** Zeit: Dienstag, 8. Juni 2021, 14:00 Uhr Ort: https://rwth.zoom.us/j/98025555800?pwd=SUgxYWR3RGxnaGNMZGg5bmNqeGZYQT09 Meeting-ID: 980 2555 5800 Kenncode: 419493 ********************************************************************** Vortragender: Holger H. Hoos (Universität Leiden) Titel: Cooperative competition: A new way of solving computationally challenging problems in AI and beyond Abstract: Progress in solving challenging problems in artificial intelligence, computer science at large, and beyond is driven, to a significant extent, by competition - regular algorithm competitions as well as comparative performance evaluation against state-of-the-art methods from the literature. A prominent example for this is the satisfiability problem in propositional logic (SAT), an NP-hard problem that not only lies at the foundations of computer science, but also plays a key role in many real-world applications, notably in ensuring the correctness of hard- and software. In this presentation, I will argue that it is time to rethink the way we assess the state of the art in solving problems such as SAT and the incentives for improving it. I will demonstrate how automated algorithm selection and configuration techniques based on sophisticated machine learning and optimisation methods have fundamentally changed not only the state of the art in solving SAT and many other NP-hard problems, but also provide a natural basis for cooperative competition - a new approach for achieving and assessing progress not merely in solving these problems, but also in the way we approach them as a scientific community. Es laden ein: die Dozentinnen und Dozenten der Informatik

3 years, 3 months

Einladung: Informatik-Oberseminar Bastian Küppers

by Küppers, Bastian

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Montag, 28. Juni 2021, 14.00 Uhr Ort: https://rwth.zoom.us/my/b.kueppers?pwd=dDFyZkJvUDRFbFlFbnh0R3hQZVVRdz09 Referent: Bastian Küppers, M.Sc. IT Center / Learning Technologies Research Group Thema: Development of a Framework for E-Assessment on Students’ Devices Abstract: In line with the general trend towards digitization, teaching at German universities is increasingly relying on digital elements, such as learning management systems or smartphone apps. This evolution is also taking place in exercises and practical courses associated with lectures. For example, it is quite common for students to use their own devices in programming exercises. However, examinations are not yet part of this development. The decision to continue writing exams on paper is often made based on reservations, most of which concern the fairness and reliability of e-assessment. In particular, students’ reservations, because they may feel responsible for a functioning device, are a major obstacle to the introduction of e-Assessment. It is therefore important to overcome these reservations, to be able to successfully implement e-assessment. In addition to reservations, financial reasons are often an obstacle. The acquisition and administration of a suitable IT infrastructure is expensive, both in terms of material costs and personnel costs. Since the majority of students already own devices that are potentially suitable for e-assessment, Bring Your Own Device (BYOD) is a possible solution to the financial aspect. However, a BYOD approach to e-assessment comes with new challenges that have to be tackled. On the way to a solution that works in the previously described scenario, requirements engineering has been a vital part of the process. Therefore, students and teachers as well as official policies have been consulted to derive the requirements to a feasible BYOD approach to e-assessment. In addition to the found requirements, a threat model has been developed to identify additional requirements to the security of such an approach. Afterwards, a software framework was developed and implemented which fulfilled the gathered requirements. Finally, the software prototype was evaluated regarding functionality, usability, performance, and security. Beyond the software prototype, an organizational framework has been developed which covers (hardware) requirements for the institute of higher education as well as important organizational details for the conduction of electronic assessment. In this talk, we discuss important key points of the research process and present the results of our work. Es laden ein: die Dozentinnen und Dozenten der Informatik

3 years, 3 months

Einladung: Informatik-Oberseminar Marcel Hark

by Marcel Hark

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Freitag, 18. Juni 2021, 15.00 Uhr Ort: Zoom-Videokonferenz Link: https://rwth.zoom.us/j/94806153993?pwd=dUIwQWYxNUx1cG1jOEpLc25WdmJwUT09 Referent: Marcel Hark M.Sc. Lehr- und Forschungsgebiet Informatik 2 Thema: Towards Complete Methods for Automatic Complexity and Termination Analysis of (Probabilistic) Programs Abstract: The increasing importance of computer programs in our everyday life has led to more and more complex software systems. To prove correctness of such a system, formal verification is the standard methodology. Two of the most important properties of a program are its termination behavior and its efficiency. Moreover, in recent years randomization in programming has gained a lot of interest. For example, to model non-deterministic behavior in real-world applications, probabilistic concepts have proved very successful. In this talk, we investigate formal verification of programs which also feature assignments via discrete probability distributions. In particular, we are interested in proving (non-)termination and inferring bounds on the (expected) worst-case runtime of such programs. In general, formal verification of programs is undecidable. Still, whenever possible, complete approaches for verifying certain properties on (sub-)classes of programs are preferable to incomplete ones since they always yield definite results, i.e., either a proof or a counterexample. Hence, we also characterize sub-classes of programs for which we can present complete approaches for analyzing termination and runtime complexity. To analyze systems arising from real-world applications, formal verification has to proceed automatically. Thus, we discuss the automation of our results as well. Es laden ein: die Dozentinnen und Dozenten der Informatik -- Marcel Hark Research Group Computer Science 2 RWTH Aachen University Ahornstr. 55 52074 Aachen Germany E-Mail: marcel.hark(a)cs.rwth-aachen.de" Phone: +49-241/80-21218 Fax: +49-241/80-22217 Room: 4208

3 years, 3 months

Einladung: Informatik-Oberseminar Philipp Wacker

by Philipp Wacker

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Mittwoch, 26. Mai 2021, 14:00 Uhr Zoom: https://rwth.zoom.us/j/92362727123?pwd=MjhBT2VaczI0RDE1c2VNQkxGckw3UT09 <https://rwth.zoom.us/j/92362727123?pwd=MjhBT2VaczI0RDE1c2VNQkxGckw3UT09> Referent: Philipp Wacker, M.Sc. Lehrstuhl Informatik 10 Thema: Interaction Techniques for Mid-Air Pen Input in Handheld Augmented Reality Abstract: Augmented Reality changes the way we interact with virtual information. Currently, virtual information is shown on 2D screens, separated from the real world. With Augmented Reality, virtual content can be shown directly embedded in the real world. This opens up the area of situated modeling in which virtual models are designed in context of the real world to, for example, print them out using a 3D printer. In an initial study, we show that sketching on physical objects improves stroke accuracy compared to strokes on virtual objects, and that features guiding a stroke, either through a concave or convex shape or through a visual guide, further improve the accuracy especially for physical objects. The most available form of Augmented Reality (AR) is Handheld Augmented Reality which shows the virtual information embedded in the camera view of everyday smartphones or tablets. However, continuously specifying a 3D position—needed, e.g., for drawing in mid-air—is not directly possible in today’s systems. We build the ARPen system to allow for situated modeling in Handheld AR, requiring only a 3D-printed pen and a consumer smartphone. But many essential interactions are not yet clear for such a bimanual system. We design and evaluate selection & manipulation techniques to adjust the pose of a mid-air object, as well as menu techniques to control properties of objects in the scene. We show that ray-casting techniques, especially through the tip of the pen, generally perform well. However, interacting on the touchscreen or even combinations of both touchscreen and mid-air input also achieve promising results. To overcome perception issues of determining the depth of virtual objects in Handheld AR, we design depth visualizations that show the position of the pen tip in relation to other objects in the scene. We identify that a heatmap visualization, coloring every object in the scene depending on their distance to the pen tip, achieves best results and was preferred by study participants. We release the ARPen system as an open-source toolbox, enabling researchers to implement and evaluate interaction techniques for Handheld AR with a mid-air pen. Our findings on essential interaction techniques provide a starting point for the development and evaluation of specialized application scenarios. Es laden ein: die Dozentinnen und Dozenten der Informatik

3 years, 4 months

2024

2023

2022

2021

2020

2019

2018

informatik-vortraege May 2021